Fig. 4: The use of iMVP to identify noise in BS-seq data.
From: Epitranscriptomic subtyping, visualization, and denoising by global motif visualization
a Cluster #1 (Type I), cluster #2 (Type II) and cluster #3 (Type III) are canonical human m5C sites. Cluster #4 is artifacts from low-complexity regions. Clusters #5 to #7 are potential false positives from Alu repeats. 2,284 sites from human tissues were analyzed. b Cluster #1 (Type I), cluster #2 (Type II) and cluster #3 (Type III) are canonical mouse m5C sites, while cluster #4 and cluster #5 are potential false positives. 2,498 sites from mouse tissues were analyzed. c IGV browser view of selected human sites in clusters #4 to #7. Samples: human heart, GSM3462633; human muscle, GSM3462639. d IGV browser view of selected mouse sites in clusters #4 and #5. Samples: mouse lung, GSM3462647; mouse testis, GSM2461443.
